Re: $GPGGA / increased number of satellites

[Gary E. Miller]

Yo Hans!

On Sat, 11 May 2024 22:55:41 +0200
Hans Mayer <gpsd@ma.yer.at> wrote:

> On 05.05.24 21:15, Gary E. Miller wrote:
> 
> Yo Gary,
> > Please post full captures.  
> 
> What do you mean with full log ?

At a minimum, I want to run it through gpsdecode and gpsfake, so I
can see why gpsd is doing what it is doing.

> I am running
> 
> gpspipe --nmea -u | grep --line-buffered GPGGA >>
> /var/log/gpsd/nmea.log &

Uh, no.  I much prefer: gpspipe -R -x 30 > raw.log

The sat data is not just in GPGGA, and you might be getting some
binary messages as well.

> Do you want to have the log of May 4th ?

If what you see it not repeatable, then your issue is moot.  So log
of any day is fine.

> Now I realized that the logs before the upgrade had 86400 lines +/-
> 1. This is one line per second. Now the files have about 10800 lines.
> Each night I do a logrotate.

One line a seconds???  No way that is enough!

> > But, beleive the the new gpsd got smarte at decoding the data that
> > you always had.  
> 
> I am quite sure that gpsd is smart, no discussion.

Typo: I meant "smarter", not "smarte" or smart. i.e. does things better
now, but maybe some worse with the better.

> 
> > What you did not notice is that the number of satellites did not
> > increase, but the number of satellite/signal pairs increased because
> > u-blox had failed to document that until recently.  
> I didn't update ZED-F9P
> > By signal, I mean: L1, L2, L5, B2, etc.
> >
> > You can see this using cgps, or xgps.
> >  
> Hmm ? Not sure what you mean. When I start xgps I don't see the bands.

Yes, you do, but not obvious.  Hover your mouse over a sat icon in 
the skyview.

> What I see is a flickering screen which changes about twice per
> second.

And that fkilering screen is chock full of data.

> On the bottom right corner in one screen there is sat seen: 50 and
> used 31, the other one say seen: 78 and used 71

And both can be true.

> When I run gpspipe ( as described above ) I get currently about 70 
> satellites in the GPGGA record.

It goes uch faster when I can see what you see.  By that, I mean send
me your raw capture.

> But when I use gpscsv to filter out the unique satellites I get about
> 49 ( in the moment )

Which comes back to what is a "satellite"?  Is a "satellite" a orbitting
object at that elevation, azimuth and pseudorange?  Or is it several
satellite/ginal pairs.  Like PRN 20/L1, PRN20/L2 and PRN 20/L5?

NMEA uses both definitions.  Very confusing.

> 
> This is the way I have done it:
> 
> mayer# gpscsv -n 10 -c SAT --header 0 | awk -F, '
> BEGIN {
>    GNSSID=99999 ;
>    SVID=99999 ;
>    PRN=99999 ;
> }

You are ignoring the sigid.  So you will count PRN 20/L1 and PRN 20/Las
as one "satellite", not two.

> But in the past ( the time before the upgrade ) the number of both 
> methods did match.

Yes, because only recently have consumer GNSS receivers received more
than one signal from a satellite, then later NMEA keeps changing their
mind how to handle that, and then only "recently" has gpsd caught up with
the secret NMEA changes of the last few years.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        gem@rellim.com  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

pgprVgGm9oz6T.pgp
Description: OpenPGP digital signature